THS4551DGKEVM: MFB bandpass filter with THS4551 is producing oscillation.

Not enough to go on, need a schematic or even better a TINA file. 

Odd that you feel the oscillation is at the BP frequency - perhaps just noise, you seem to have a lot of gain here. 

Thanks for your reply.

My circuit order is following. I measure/see signals output of VGA by test points. 

signal input   ->  preamplifier (THS4551)(gain =20dB)    ->   BPF(THS4551) (gain=1)   ->  VGA  (0-80dB)(AD8338) -> SCOP.

VGA doesn't produce 70khz so it can just come by BPF. 

I added tina file that just includes filter circuit. There is not continuous oscillation in tina. I seem just impulse response ( with 1khz square impulse) but the oscillation is continuous in real circuit at 40-60dB  gain ( with VGA) after filter.

For example, if I give 80-100dB of gain  to output of any BPF while filter input connected to signal ground, can I see oscillation at center frq at output of VGA?

Or  Is there any noise (or ripple) at the central frequency at the band pass filter output as a few microvolt?

1643.BPF.TSC.

Thanks,

That is quite the 8th order BP, if Ijust run a spot noise sim, it builds up to 220nV at the final output, if you really have the VGA set to a gain of 80dB, that is 10000*220nV = 2.2mV spot noise. 

In any case, whenever you have these cascaded high gain stages, I like (others prefer a star) a cascaded pi filter on the supply going backwards from the output to input to break supply loops. 

Thanks a lot ,

Sorry, I didn't understand exactly,  Please can you give me more detail with a schematic about your pi type filter?  ( if possible )

and How did you find 220nV?

oops I forgot to put the spot noise sim in, also, the pi filter thing is the last example in this 1990 app note I wrote, 

http://www.ti.com/lit/an/snoa367c/snoa367c.pdf

Thanks a lot,

Please can you share tina file?

It is your file, I just ran a spot noise sim on it. 

Hi again,

I perform FFT  by scop at output of VGA . There are band pass frequencies, not just center frequency (70Khz).

My circuit includes two VGA as cascaded. Circuit Order :

signal input -> preamplifier( gain =1, single ended to differential) ->  bandpass  -> VGA1(0-80db)  ->VGA2(0-80db) -> ADC

if my current problem:

if VGA1 gain is increased, bandpass frequencies become clear and they are increased in dBV (in FFT) by VGA1 gain.(VGA2 = 1 or different)

But if VGA1 gain equals 1 and  if VGA2 gain is increased, bandpass frequencies aren't increased in total band in FFT,  total noise(all bands in FFT) is just increased by VGA2 gain.

How this can occur ?

thanks a lot